\n
\n "},"metadata":{},"execution_count":16}],"execution_count":16},{"cell_type":"markdown","source":"REGRESIÓN LINEAL MÚLTIPLE","metadata":{"id":"uKXLvubm0MNV","cell_id":"ed2651c8d37f43fdad5310ab7bb0b9d7","deepnote_cell_type":"markdown"}},{"cell_type":"code","source":"dataset = pd.read_csv(\"petrol_consumption.csv\", sep = \",\")","metadata":{"id":"O4Jvlbtf0MNW","cell_id":"ee091d17f26246a1862091f3f5347944","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":628,"user_tz":180,"timestamp":1650392547720},"deepnote_cell_type":"code"},"outputs":[],"execution_count":17},{"cell_type":"code","source":"#Vemos el head\ndataset.head()","metadata":{"id":"jhuZhjUH0MNW","colab":{"height":206,"base_uri":"https://localhost:8080/"},"cell_id":"4eb8489816274860935aeb5d71bd2c9f","outputId":"6a7b753e-524d-43cc-8cef-4dfc21f08da1","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":9,"user_tz":180,"timestamp":1650392548917},"deepnote_cell_type":"code"},"outputs":[{"output_type":"execute_result","data":{"text/plain":" Petrol_tax Average_income Paved_Highways Population_Driver_licence(%) \\\n0 9.0 3571 1976 0.525 \n1 9.0 4092 1250 0.572 \n2 9.0 3865 1586 0.580 \n3 7.5 4870 2351 0.529 \n4 8.0 4399 431 0.544 \n\n Petrol_Consumption \n0 541 \n1 524 \n2 561 \n3 414 \n4 410 ","text/html":"\n \n
\n \n\n\n \n
\n
\n \n \n \n\n \n | \n | Actual | \nPredicted | \nSesgo | \nError_porc | \n
|---|---|---|---|---|
| 0 | \n81 | \n83.188141 | \n-2.188141 | \n-2.701409 | \n
| 1 | \n30 | \n27.032088 | \n2.967912 | \n9.893041 | \n
| 2 | \n21 | \n27.032088 | \n-6.032088 | \n-28.724227 | \n
| 3 | \n76 | \n69.633232 | \n6.366768 | \n8.377327 | \n
| 4 | \n62 | \n59.951153 | \n2.048847 | \n3.304591 | \n
\n
\n "},"metadata":{},"execution_count":18}],"execution_count":18},{"cell_type":"code","source":"dataset.shape","metadata":{"id":"UHeQoz2dv9oZ","colab":{"base_uri":"https://localhost:8080/"},"cell_id":"c1b1da7d98f145fa9453ce2ad82d39a3","outputId":"86269dff-de3a-439b-a50a-051a56f5895a","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":4,"user_tz":180,"timestamp":1650392550255},"deepnote_cell_type":"code"},"outputs":[{"output_type":"execute_result","data":{"text/plain":"(48, 5)"},"metadata":{},"execution_count":19}],"execution_count":19},{"cell_type":"code","source":"#Estadisticas\ndataset.describe()","metadata":{"id":"3uzSWdsR0MNW","colab":{"height":300,"base_uri":"https://localhost:8080/"},"cell_id":"a505398abfac4e89be2a23f6759abd55","outputId":"674a0124-1aab-43b3-e9af-685d5707f90e","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":368,"user_tz":180,"timestamp":1650392551077},"deepnote_cell_type":"code"},"outputs":[{"output_type":"execute_result","data":{"text/plain":" Petrol_tax Average_income Paved_Highways \\\ncount 48.000000 48.000000 48.000000 \nmean 7.668333 4241.833333 5565.416667 \nstd 0.950770 573.623768 3491.507166 \nmin 5.000000 3063.000000 431.000000 \n25% 7.000000 3739.000000 3110.250000 \n50% 7.500000 4298.000000 4735.500000 \n75% 8.125000 4578.750000 7156.000000 \nmax 10.000000 5342.000000 17782.000000 \n\n Population_Driver_licence(%) Petrol_Consumption \ncount 48.000000 48.000000 \nmean 0.570333 576.770833 \nstd 0.055470 111.885816 \nmin 0.451000 344.000000 \n25% 0.529750 509.500000 \n50% 0.564500 568.500000 \n75% 0.595250 632.750000 \nmax 0.724000 968.000000 ","text/html":"\n \n
\n \n\n\n \n
\n
\n \n \n \n\n \n | \n | Petrol_tax | \nAverage_income | \nPaved_Highways | \nPopulation_Driver_licence(%) | \nPetrol_Consumption | \n
|---|---|---|---|---|---|
| 0 | \n9.0 | \n3571 | \n1976 | \n0.525 | \n541 | \n
| 1 | \n9.0 | \n4092 | \n1250 | \n0.572 | \n524 | \n
| 2 | \n9.0 | \n3865 | \n1586 | \n0.580 | \n561 | \n
| 3 | \n7.5 | \n4870 | \n2351 | \n0.529 | \n414 | \n
| 4 | \n8.0 | \n4399 | \n431 | \n0.544 | \n410 | \n
\n
\n "},"metadata":{},"execution_count":20}],"execution_count":20},{"cell_type":"code","source":"# 1 )Preparación de datos\nX = dataset[['Petrol_tax', 'Average_income', 'Paved_Highways','Population_Driver_licence(%)']]\ny = dataset['Petrol_Consumption']","metadata":{"id":"4ADlRF6t0MNX","cell_id":"a843e49a107946d99333e9a2f54ce304","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":279,"user_tz":180,"timestamp":1650392553127},"deepnote_cell_type":"code"},"outputs":[],"execution_count":21},{"cell_type":"code","source":"# 2) Separacion en train y test\nfrom sklearn.model_selection import train_test_split\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)","metadata":{"id":"k25OJEmM0MNX","cell_id":"3e80bdc8bb97405f8d5baf189b0b7e66","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":3,"user_tz":180,"timestamp":1650392553542},"deepnote_cell_type":"code"},"outputs":[],"execution_count":22},{"cell_type":"code","source":"#Entrenamiento del modelo\nfrom sklearn.linear_model import LinearRegression\nregressor = LinearRegression()\nregressor.fit(X_train, y_train)","metadata":{"id":"mtIo08po0MNX","colab":{"base_uri":"https://localhost:8080/"},"cell_id":"f1866e611e4145d0a969af3a9042e358","outputId":"1f7732f5-0fe5-4d27-83e6-8007717c1482","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":273,"user_tz":180,"timestamp":1650392555650},"deepnote_cell_type":"code"},"outputs":[{"output_type":"execute_result","data":{"text/plain":"LinearRegression()"},"metadata":{},"execution_count":23}],"execution_count":23},{"cell_type":"markdown","source":"Como se dijo anteriormente, en caso de regresión lineal multivariable, el modelo de regresión tiene que encontrar los coeficientes más óptimos para todos los atributos. Para ver qué coeficientes ha elegido nuestro modelo de regresión, podemos ejecutar el siguiente script:","metadata":{"id":"DVHaO8yu0MNY","cell_id":"41a7e8922c94492d811737e53610560b","deepnote_cell_type":"markdown"}},{"cell_type":"code","source":"# 'Petrol_tax', 'Average_income', 'Paved_Highways','Population_Driver_licence(%)'\nregressor.coef_","metadata":{"id":"wQ_jirk1xBY5","colab":{"base_uri":"https://localhost:8080/"},"cell_id":"4ba886c15d284b019efce74e48ecdb64","outputId":"af9827f5-5f93-4c80-9002-b15dbdd72102","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":7,"user_tz":180,"timestamp":1650392556488},"deepnote_cell_type":"code"},"outputs":[{"output_type":"execute_result","data":{"text/plain":"array([-3.69937459e+01, -5.65355145e-02, -4.38217137e-03, 1.34686930e+03])"},"metadata":{},"execution_count":24}],"execution_count":24},{"cell_type":"code","source":"regressor.intercept_","metadata":{"id":"wBh7MxHRxU9k","colab":{"base_uri":"https://localhost:8080/"},"cell_id":"76c9e413864b4cbdbfb82375ec2cec27","outputId":"f24b60eb-f018-40c0-d93c-6a08c5e21b22","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":5,"user_tz":180,"timestamp":1650392558061},"deepnote_cell_type":"code"},"outputs":[{"output_type":"execute_result","data":{"text/plain":"361.45087906653225"},"metadata":{},"execution_count":25}],"execution_count":25},{"cell_type":"code","source":"X.columns","metadata":{"id":"YdtlK7FDyiHd","colab":{"base_uri":"https://localhost:8080/"},"cell_id":"5dbd800f7dea4839bec200fabf0b8be9","outputId":"6e1faad1-449c-482a-da59-acb808f56652","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":6,"user_tz":180,"timestamp":1650392559150},"deepnote_cell_type":"code"},"outputs":[{"output_type":"execute_result","data":{"text/plain":"Index(['Petrol_tax', 'Average_income', 'Paved_Highways',\n 'Population_Driver_licence(%)'],\n dtype='object')"},"metadata":{},"execution_count":26}],"execution_count":26},{"cell_type":"code","source":"coeff_df = pd.DataFrame(regressor.coef_, X.columns, columns=['Coefficient'])\ncoeff_df","metadata":{"id":"OU4eIQiA0MNY","colab":{"height":175,"base_uri":"https://localhost:8080/"},"cell_id":"8da275b8b505484ca68d28b6b069c506","outputId":"931e5b2b-81d9-4c5b-a46f-0f3205187018","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":8,"user_tz":180,"timestamp":1650392559690},"deepnote_cell_type":"code"},"outputs":[{"output_type":"execute_result","data":{"text/plain":" Coefficient\nPetrol_tax -36.993746\nAverage_income -0.056536\nPaved_Highways -0.004382\nPopulation_Driver_licence(%) 1346.869298","text/html":"\n \n
\n \n\n\n \n
\n
\n \n \n \n\n \n | \n | Petrol_tax | \nAverage_income | \nPaved_Highways | \nPopulation_Driver_licence(%) | \nPetrol_Consumption | \n
|---|---|---|---|---|---|
| count | \n48.000000 | \n48.000000 | \n48.000000 | \n48.000000 | \n48.000000 | \n
| mean | \n7.668333 | \n4241.833333 | \n5565.416667 | \n0.570333 | \n576.770833 | \n
| std | \n0.950770 | \n573.623768 | \n3491.507166 | \n0.055470 | \n111.885816 | \n
| min | \n5.000000 | \n3063.000000 | \n431.000000 | \n0.451000 | \n344.000000 | \n
| 25% | \n7.000000 | \n3739.000000 | \n3110.250000 | \n0.529750 | \n509.500000 | \n
| 50% | \n7.500000 | \n4298.000000 | \n4735.500000 | \n0.564500 | \n568.500000 | \n
| 75% | \n8.125000 | \n4578.750000 | \n7156.000000 | \n0.595250 | \n632.750000 | \n
| max | \n10.000000 | \n5342.000000 | \n17782.000000 | \n0.724000 | \n968.000000 | \n
\n
\n "},"metadata":{},"execution_count":27}],"execution_count":27},{"cell_type":"code","source":"#Realizando las predicciones\ny_pred = regressor.predict(X_test)\ny_pred","metadata":{"id":"ddBKEWyr0MNZ","colab":{"base_uri":"https://localhost:8080/"},"cell_id":"07ad97d5c69445bb9d2d3a881d20a654","outputId":"c587e788-c260-43e7-8c38-bb281033feec","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":297,"user_tz":180,"timestamp":1650392563387},"deepnote_cell_type":"code"},"outputs":[{"output_type":"execute_result","data":{"text/plain":"array([606.69266519, 673.77944169, 584.99149034, 563.53691024,\n 519.05867235, 643.46100256, 572.89761422, 687.07703573,\n 547.6093662 , 530.03762971])"},"metadata":{},"execution_count":28}],"execution_count":28},{"cell_type":"code","source":"y_test","metadata":{"id":"zbTwKgGuzBJC","colab":{"base_uri":"https://localhost:8080/"},"cell_id":"5a259d544fe843309d4faa7d30b1811b","outputId":"652d6363-07ea-48cb-90cd-05afc320c4ae","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":337,"user_tz":180,"timestamp":1650392564974},"deepnote_cell_type":"code"},"outputs":[{"output_type":"execute_result","data":{"text/plain":"27 631\n40 587\n26 577\n43 591\n24 460\n37 704\n12 525\n19 640\n4 410\n25 566\nName: Petrol_Consumption, dtype: int64"},"metadata":{},"execution_count":29}],"execution_count":29},{"cell_type":"markdown","source":"Para comparar los valores de salida reales X_test con los valores predichos, convertimos en df:","metadata":{"id":"AOslg9Xl0MNZ","cell_id":"8ea147c6ccba4276b8271700b27540c4","deepnote_cell_type":"markdown"}},{"cell_type":"code","source":"df = pd.DataFrame({'Actual': y_test, 'Predicted': y_pred})\ndf['Sesgo']=df.Actual -df.Predicted\ndf['Error_porc']=((df.Actual -df.Predicted)/df.Actual) *100\ndf","metadata":{"id":"nLFDzsL6y8A_","colab":{"height":362,"base_uri":"https://localhost:8080/"},"cell_id":"d59c89d738784d9aa362ff35a02cafa8","outputId":"b6278d81-1fec-464c-c9bc-0c1433b693d5","executionInfo":{"user":{"userId":"09471607480253994520","displayName":"David Francisco Bustos Usta"},"status":"ok","elapsed":286,"user_tz":180,"timestamp":1650392567166},"deepnote_cell_type":"code"},"outputs":[{"output_type":"execute_result","data":{"text/plain":" Actual Predicted Sesgo Error_porc\n27 631 606.692665 24.307335 3.852193\n40 587 673.779442 -86.779442 -14.783551\n26 577 584.991490 -7.991490 -1.385007\n43 591 563.536910 27.463090 4.646885\n24 460 519.058672 -59.058672 -12.838842\n37 704 643.461003 60.538997 8.599289\n12 525 572.897614 -47.897614 -9.123355\n19 640 687.077036 -47.077036 -7.355787\n4 410 547.609366 -137.609366 -33.563260\n25 566 530.037630 35.962370 6.353776","text/html":"\n \n
\n \n\n\n \n
\n
\n \n \n \n\n \n | \n | Coefficient | \n
|---|---|
| Petrol_tax | \n-36.993746 | \n
| Average_income | \n-0.056536 | \n
| Paved_Highways | \n-0.004382 | \n
| Population_Driver_licence(%) | \n1346.869298 | \n
\n
\n "},"metadata":{},"execution_count":30}],"execution_count":30},{"cell_type":"markdown","source":"\n\n
\n \n\n\n \n
\n
\n \n \n \n\n \n | \n | Actual | \nPredicted | \nSesgo | \nError_porc | \n
|---|---|---|---|---|
| 27 | \n631 | \n606.692665 | \n24.307335 | \n3.852193 | \n
| 40 | \n587 | \n673.779442 | \n-86.779442 | \n-14.783551 | \n
| 26 | \n577 | \n584.991490 | \n-7.991490 | \n-1.385007 | \n
| 43 | \n591 | \n563.536910 | \n27.463090 | \n4.646885 | \n
| 24 | \n460 | \n519.058672 | \n-59.058672 | \n-12.838842 | \n
| 37 | \n704 | \n643.461003 | \n60.538997 | \n8.599289 | \n
| 12 | \n525 | \n572.897614 | \n-47.897614 | \n-9.123355 | \n
| 19 | \n640 | \n687.077036 | \n-47.077036 | \n-7.355787 | \n
| 4 | \n410 | \n547.609366 | \n-137.609366 | \n-33.563260 | \n
| 25 | \n566 | \n530.037630 | \n35.962370 | \n6.353776 | \n
\nCreated in Deepnote","metadata":{"created_in_deepnote_cell":true,"deepnote_cell_type":"markdown"}}],"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"Ejemplo 2-Regresión.ipynb","provenance":[],"collapsed_sections":[]},"deepnote":{},"kernelspec":{"name":"python3","language":"python","display_name":"Python 3"},"varInspector":{"cols":{"lenVar":40,"lenName":16,"lenType":16},"kernels_config":{"r":{"library":"var_list.r","varRefreshCmd":"cat(var_dic_list()) ","delete_cmd_prefix":"rm(","delete_cmd_postfix":") "},"python":{"library":"var_list.py","varRefreshCmd":"print(var_dic_list())","delete_cmd_prefix":"del ","delete_cmd_postfix":""}},"window_display":false,"types_to_exclude":["module","function","builtin_function_or_method","instance","_Feature"]},"language_info":{"name":"python","version":"3.8.5","mimetype":"text/x-python","file_extension":".py","pygments_lexer":"ipython3","codemirror_mode":{"name":"ipython","version":3},"nbconvert_exporter":"python"},"deepnote_notebook_id":"c88c8868da2348beb5f021296eb77089","deepnote_execution_queue":[]}}